In many cryogenic applications components, e.g. superconducting coils for magnetic resonance imaging (MRI), superconducting transformers, generators, electronics, are cooled by keeping them in contact with a volume of liquefied gases (e.g. Helium, Neon, Nitrogen, Argon, Methane). Any dissipation in the components or heat getting into the system causes the volume to part boil off. To account for the losses, replenishment is required. This service operation is considered to be problematic by many users and great efforts have been made over the years to introduce refrigerators that recondense any lost liquid right back into the bath.
As an example of prior art, an embodiment of a two stage Gifford McMahon (GM) coldhead recondenser of an MRI magnet is shown in FIG. 1. In order for the GM coldhead, indicated generally by 10, to be removable for service or repair, it is inserted into a sock, which connects the outside face of a vacuum vessel 16 (at room temperature) to a helium bath 18 at 4K. MRI magnets are indicated at 20. The sock is made of thin walled stainless steel tubes forming a first stage sleeve 12, and a second stage sleeve 14 in order to minimise heat conduction from room temperature to the cold end of the sock operating at cryogenic temperatures. The sock is filled with helium gas 30, which is at about 4.2 K at the cold end and at room temperature at the warm end. The first stage sleeve 12 of the coldhead is connected to an intermediate heat station of the sock 22, in order to extract heat at an intermediate temperature, e.g. 40K-80 K, and to which sleeve 14 is also connected. The second stage of the coldhead 24 is connected to a helium gas recondenser 26. Heat arises from conduction of heat down through the neck, heat radiated from a thermal radiation shield 42 as well as any other sources of heat for example, from a mechanical suspension system for the magnet, (not shown) and from a service neck (also not shown) used for filling the bath with liquids, instrumentation wiring access, gas escape route etc. The intermediate section 22 shows a passage 38 to enable helium gas to flow from the volume encircled by sleeve 14. A number of passages may be annularly distributed about the intermediate section. The latter volume is also in fluid connection with the main bath 18 in which the magnet 20 is placed. Also shown is a flange 40 associated with sleeve 12 to assist in attaching the sock to the vacuum vessel 16. A radiation shield 42 is placed intermediate the helium bath and the wall of the outer vacuum vessel.
The second stage of the coldhead is acting as a recondensor at about 4.2 K. As it is slightly colder than the surrounding He gas, gas is condensed on the surface (which can be equipped with fins to increase surface area) and is dripped back into the liquid reservoir. Condensation locally reduces pressure, which pulls more gas towards the second stage. It has been calculated that there are hardly any losses due to natural convection of Helium, which has been verified experimentally provided that the coldhead and the sock are vertically oriented (defined as the warm end pointing upwards). Any small differences in the temperature profiles of the Gifford McMahon cooler and the walls would set up gravity assisted gas convection, as the density change of gas with temperature is great (e.g. at 4.2. K the density is 16 kg/m3; at 300 K the density is 0.16 kg/m3). Convection tends to equilibrate the temperature profiles of the sock wall and the refrigerator. The residual heat losses are small.
When the arrangement is tilted, natural convection sets up huge losses. A solution to this problem has been described in U.S. patent, U.S. Pat. No. 5,583,472, to Mitsubishi. Nevertheless, this will not be further discussed here, as this document relates to arrangements which are vertically oriented or at small angles (<30°) to the vertical.
It has been shown that Pulse Tube Refrigerators (PTRs) can achieve useful cooling at temperatures of 4.2 K (the boiling point of liquid helium at normal pressure) and below (C. Wang and P. E. Gifford, Advances in Cryogenic Engineering, 45, Edited by Shu et a., Kluwer Academic/Plenum Publishers, 2000, pp. 1-7). Pulse tube refrigerators are attractive, because they avoid any moving parts in the cold part of the refrigerator, thus reducing vibrations and wear of the refrigerator. Referring now to FIG. 2, there is shown a PTR 50 comprising an arrangement of separate tubes, which are joined together at heat stations. There is one regenerator tube 52, 54 per stage, which is filled with solid materials in different forms (e.g. meshes, packed spheres, powders). The materials act as a heat buffer and exchange heat with the working fluid of the PTR (usually He gas at a pressure of 1.5-2.5 MPa). There is one pulse tube 56, 58 per stage, which is hollow and used for expansion and compression of the working fluid. In two stage PTRs, the second stage pulse tube 56 usually links the second stage 60 with the warm end 62 at room temperature, the first stage pulse tube 58 linking the first stage 64 with the warm end.
It has been found, that PTRs operating in vacuum under optimum conditions usually develop temperature profiles along the length of the tubes that are significantly different one tube to another in the same temperature range and also from what would be a steady state temperature profile in a sock. This is shown in FIG. 3.
Another prior art pulse tube refrigerator arrangement is shown in FIG. 4 wherein a pulse tube is inserted into a sock, and is exposed to a helium atmosphere wherein gravity induced-convection currents 70, 72 are set up in the first and second stages. The PTR unit 50 is provided with cold stages 31, 33 which are set in a recess in an outer vacuum container 16. A radiation shield 42 is provided which is in thermal contact with first sleeve end 22. A recondenser 26 is shown on the end wall of second stage 33. If at a given height the temperatures of the different components are not equal, the warmer components will heat the surrounding helium, giving it buoyancy to rise, while at the colder components the gas is cooled and drops down. The resulting thermal losses are huge, as the density difference of helium gas at 1 bar changes by a factor of about 100 between 4.2 K and 300 K. The net cooling power of a PTR might be e.g. 40 W at 50 K, and 0.5 W to 1 W at 4.2 K. The losses have been calculated to be of the order of 5-20 W. The internal working process of a pulse tube will, in general, be affected although this is not encountered in GM refrigerators. In a PTR, the optimum temperature profile in the tubes, which is a basis for optimum performance, arises through a delicate process balancing the influences of many parameters, e.g. geometries of all tubes, flow resistivities, velocities, heat transfer coefficients, valve settings etc. (A description can be found in Ray Radebaugh, proceedings of the 6th International Cryogenic Engineering Conference, Kitakkyushu, Japan, 20-24 May, 1996, pp. 22-44).
Therefore, in a helium environment, PTRs do not necessarily reach temperatures of 4 K, although they are capable of doing so in vacuum. Nevertheless, if the PTR is inserted in a vacuum sock with a heat contact to 4 K through a solid wall, it would work normally. Such a solution has been described for a GM refrigerator (U.S. patent U.S. Pat. No. 5,613,367 to William E. Chen, G E) although the use of a PTR would be possible and be straightforward. The disadvantage, however, is that the thermal contact of the coldhead at 4 K would produce a thermal impedance, which effectively reduces the available power for refrigeration. As an example, With a state of the art thermal joint made from an Indium washer, a thermal contact resistance of 0.5 K/W can be achieved at 4 K (see e.g. U.S. Pat. No. 5,918,470 to GE). If a cryocooler can absorb 1 W at 4.2 K (e.g. the model RDK 408 by Sumitomo Heavy Industries) then the temperature of the recondensor would rise to 4.7 K, which would reduce the current carrying capability of the superconducting wire drastically. Alternatively, a stronger cryocooler would be required to produce 1 W at 3.7 K initially to make the cooling power available on the far side of the joint.
FIG. 5 shows an example of such a PTR arrangement 76. The component features are substantially the same as shown in FIG. 4. Thermal washer 78 is provided between the second stage of the PTR coldhead and a finned heat sink 26. A helium-tight wall is provided between the thermal washer and the heat sink.